Characterization of a structurally tricistronic gene of human cytomegalovirus composed of U(s)18, U(s)19, and U(s)20.

(1)

JOURNAL OFVIROLOGY,Apr. 1993,p. 2043-2054 0022-538X/93/042043-12$02.00/0

Characterization of

a

Structurally

Tricistronic

Gene of

Human

Cytomegalovirus Composed of

Us18,

Us19,

and

Us20

YAW-WENGUO' ANDENG-SHANG HUANG'2,3,4*

Curiculumof Genetics, 1 Department of Medicine,2Departmentof Microbiology and Immunology,3 andLineberger Comprehensive Cancer Center,4 University of North Carolina

atChapel Hill, Chapel Hill, North Carolina 27599-7295 Received 22October 1992/Accepted 16 December 1992

Atricistronic genemapped between 0.91 and 0.93 mapunits within theEcoRI D fragment of the human cytomegalovirus uniqueshortregion

(Us)

has beencloned, sequenced,andexpressedinvitro.Cloned cDNAs of 2.3, 1.8, and 1.1 kb derived from this region were isolated from a Agtll cDNA library made from virus-infected fibroblastsand usedfor this study. Two majorclassesof 3'.coterminalmRNAs, 2.8 and 1.1 kb, weretranscribed fromthisregion. Sequence analysis of the cDNAs andtheupstreamgenomic DNA revealed threeopenreadingframes (ORFs),

Us18, Us19,

and

Us20,

anda common polyadenylation signal located15 bases upstream of thepoly(A) tail of both the2.85- and 1.1-kb mRNAs. Protein structureanalysespredicted theexistence ofmultiple hydrophobic moieties, suggestingthat the

Us18, US19,

and

Us20

polypeptideswere transmembrane proteins. The major transcription initiation site, determined by primer extension and S1 nucleasemapping,for the 2.85-kbtranscriptwaslocatedrightatthe firstinitiationcodon ofthe

Us20

ORF. Therewas notypicalTATA boxorCAAT boxupstreamof the 2.85-kb mRNAcapsiteexceptforaTATAAGA

sequencethatwasfoundabout 210bpdownstream from themajorcapsite. The1.1-kbtranscriptwasinitiated 33 bpupstream of the

Us18

translationinitiationsite,andanatypicalTATAboxsequence (GATAAGA)was

found 22bpupstream of the transcriptionstartsite. Differences intranscriptionkinetics and sensitivities to metabolicinhibitors suggestthattheywereregulated by differentmechanisms; the 2.85-kb mRNAbelongsto theearly

(13)

class oftranscripts, while the 1.1-kb mRNA isa late('y) message. Subgenomic DNAsegments derived from the

Us18, Us19,

and

Us20

ORFsweresubcloned andoverexpressedinEscherichiacoliasfusion proteinswithglutathione-s-transferase.Western immunoblotanalysiswithantibodiesagainst the

Us18, Us19,

and

Us20

fusion proteins detectedvirus-specific polypeptides with molecular sizes of36, 32, and 43 kDa, respectively.Allthree antibodies alsoexhibitedapositiveimmunofluorescence reaction with human cytomeg-alovirus-infected cells harvested atlate stages ofinfection.

Humancytomegalovirus (HCMV)is the fourth herpesvi-rustobesequencedand has thegreatestgenomic complexity

among the human viruses discovered to date(3). The viral

genome is a double-stranded DNA 230 kb in length. More than 200 openreading frames (ORFs) can be derived from this viral genome (3). Like that ofherpes simplex viruses, the HCMVgenomeiscomposedofaunique long (UL)anda unique short

(Us)

segment, each bounded by terminal re-peatsthatpermitULand

Us

inversions to form four isomeric molecules (9,13, 14).TheHCMV

Us

regionisverydifferent from that of other human herpesviruses. The 38-kb

Us

regionof HCMV containsatleast six families ofhomologous ORFs, andexceptfor theGprotein-coupled receptor (GCR) family,noneof these showhomologytoother known human herpesviruses ORFs (25). Each family contains 2 to 13 membersorORFs whichoccureitherastandemarrays(the

Us2, Us6,

andUs12families) ordispersedin theUsregion

(the

Us1, Us22,

and GCRfamilies).

The

Us2

family, composed of the

Us2

and

US3

ORFs, produces a group of abundant immediate-early (IE) tran-scripts encoding a group ofputative membrane-associated glycoproteins (24). The US6 family is composed of six

members,

Us6

to Us11, which encode a group of early transcripts(includingtwobicistronicmRNAs)foragroup of

virion-associated envelope glycoprotein complexes, desig-natedgCII (gp47-52 [6]). TheUs22familyhas 12members,

*Correspondingauthor.

only4ofwhich,

Us22, Us23, Us24,

and

Us26,

arelocated within the

Us

region. This family encodes a group ofearly viralproteinswhichcontain N-linkedglycosylation sitesand C-terminal charged residues but no obvious hydrophobic transmembraneregion. HCMVproteinICP22wasidentified as a member of this

Us22

family. This protein is an early viralproteinpredominantlylocalized inthe nucleusbut also secreted into the culture medium from infected cells (15). The GCR family produces membrane-spanning glycopro-teins which havesequencehomologywith theopsinfamily of cellular receptors (3). All three members of the GCR family,

Us27,

Us28,

and

Us33,

have sevenpotential mem-brane-spanning regions. Members of this diverse family transduce differentsignalsinavarietyofsystemsthat have roles invision,olfaction, memory, learning, andregulation of thecirculatorysystem(16). Exceptfor data obtained from computer analysis of viral DNA sequences, there is little information about the remaining two large gene families,

Usl

and

Us12.

The

Us12

familyhas 10 members

(Us12

to

Us21),

whichareclustered andtandemly arrangedbetween

the

Us6

and

Us22

familiesin the

Us

region.Members ofthis family aregrouped together bythe characteristic ofhaving multiple hydrophobic regionswithin their products, which

are tentatively classified as membrane-associated proteins with potential glycosylation sites. Each has up to seven highly hydrophobic putative transmembrane regions. The HCMV ORFs studied in this communicationbelongto this family.

Most virus genes are monocistronic, encoding only one

2043

Vol. 67, No. 4

on November 9, 2019 by guest

http://jvi.asm.org/

(2)

polypeptideinonegene.HCMV isanexceptionstothisrule,

as polystronic messages are not uncommon in this human

herpesvirus.For example, pp65 andpp7l (UL82 andUL83)

areencoded bybicistronic genes in HCMV(3,

18).

In this

communication, we show that the

unique

short

region

of HCMV between 0.91 and 0.93 map units encodesa

structur-ally

tricistronic geneforthe

US18,

Us19,

and

Us20

polypep-tides. Transcription of polycistronic genes in prokaryotic organismsis regulated in a

sequential

manner. In general,

thecodingregion closestto the promoter

regulatory

region

produces

moreproducts.

Therefore,

atthe

beginning

of this

study,weassumed that

regulation

of thetranscriptionof this HCMVtricistronic gene might be similar to that of a

pro-karyoticpolycistronicgene.Northern

(RNA

blot) hybridiza-tion, however, revealed two

major

mRNAs with different

transcription kinetics. Therefore, the transcriptional and translationalregulationsof these messageswere major

sub-jectsforthisstudy.

Sequence analysis

of these three

nonoverlapping

ORFs

(transcription

is leftward, from

Us20

to

Us18)

predicted

three

polypeptides

of342,240,and 274 aminoacid

residues,

respectively. Toverifywhether these three ORFs are func-tional andtoconfirmthat this tricistronic gene is translated in HCMV-infected

cells,

subgenomic DNA fragments of these three ORFs were

engineered

and subcloned into a bacterial

expression

vector

containing

the

carboxy

terminus of the

glutathione-S-transferase

(GST)

gene from Schisto-somajaponicum under the control ofa tac promoter, the

pGEX-GST

system

(21).

These

bacterially expressed

fusion

proteinswere used as

antigens

to prepare

polyclonal

anti-bodies in rabbits forusein Westernimmunoblot and

immu-nofluoresence

analyses.

MATERUILSANDMETHODS

Virus, cells, andbacterial strains. Human

embryonic lung

(HEL)

fibroblasts

(ATCC

HEL

229;

passages15to

24)

were

used for

propagation

of the virus and allbiochemicalstudies. HEL cells were cultured in

Eagle's

minimal essential

me-dium

supplemented

with 10% fetal calfserum. The Towne

strain of HCMV

(passage

34to

39)

was

propagated

in HEL cells at a low

multiplicity

of infection

(0.001 PFU/cell)

for

seed

stocks,

andthe virustiterwasdeterminedbya

plaque

assayin HEL cells. Formost

experiments,

HELcellswere grownto nearconfluence inaculture flask and infected with Townestrain HCMVat a

multiplicity

of infection of

approx-imately

2

PFU/cell.

Virus absorption was carried out in a

CO2

incubator at

37°C

for 2 h. Virus-infected cells were harvestedatthetimes indicated. Two bacterial strainswere used in these

experiments.

Eschenchiacoli JM107wasused in the

propagation

of

plasmid

DNAandM13single-stranded

phage

DNA

template.

E. coliY1090wasused in the

propa-gationofAgtll cDNAclones(8).

Construction and screening ofcDNA libraries. Total and

cytoplasmic

RNAswereisolatedfromHCMV-infectedHEL

cells at 72 hpostinfection (p.i.) bytheguanidinium isothio-cyanatemethod(10).A cDNAlibrarywasconstructed from this RNAasdescribedpreviously (8). Essentially,

polyade-nylated

[poly(A)+] RNAswerepurified by using

oligo(dT)-cellulose column chromatography. Double-stranded cDNA was synthesized frompoly(A)+RNAby using oligo(dT) as

the

primer.

The cDNAsweremethylatedwith EcoRI

meth-ylase

beforethe EcoRIlinkerwasadded. After fractionation

through

aSepharose4Bcolumn, cDNAslarger than 0.5 kb

were inserted into the EcoRI site of

Xgtll

andpackaged in

vitrowithpackagingextractsobtained from Promega

(Mad-ison, Wis.)asdescribed before(8). Two cDNA libraries with 1.5 x 106 and 2 x 106 recombinants wereconstructed from total andcytoplasmic mRNA,respectively.

DNA labeling. The radioactive viral DNA probe was labeled by the random-primed DNA labeling method (Ran-dom PrimerLabeling SystemU1110;Promega). To label the 5' termini of an oligonucleotide probe, 20 pmol of dephos-phorylated oligonucleotides in an oligonucleotide kinase buffer with 0.1 mCi of [_y-32P]ATP (3,000 Ci/mmol; Amer-sham Corp.) and 4 U of T4 polynucleotide kinase was incubated at 37°C for 30min. After heat inactivation at 65°C for 10 min, the labeled oligonucleotides were purified by

three rounds of ethanolprecipitation.

Southernhybridization. Southern hybridizations were per-formed by standard methods (20). Prehybridization was carried out in a mixture containing 25 mM KPO4 (pH 7.4), 6x SSC(0.9 M NaCl, 0.09 M sodium citrate), 5xDenhardt's solution (0.1% Ficoll, 0.1%polyvinylpyrrolidone, 0.1% bo-vine serumalbumin), 50 ,ug of denatured salmon sperm DNA per ml, and0.1% sodium dodecyl sulfate (SDS) at 42°C for 6 h. After prehybridization, denatured 32P-labeled cDNA probe (5 x 105 cpm/ml)was added and incubated at 42°C overnight. Following hybridization, filters were washed twice in lx SSC-0.1% SDS and twice in 0.25x SSC-0.1% SDSat roomtemperature for 15 min each.

Northern hybridization. Total RNAs (8 ,g per well) pre-pared from mock- and HCMV-infected HEL cells by guani-diniumisothiocyanate cell lysis and cesium chloride equilib-rium centrifugation were subjected to electrophoresis in a 1.2% agarose gel containing 6% formaldehyde and then transferredtonitrocelluloseasdescribedbefore(10,20). The

resulting filterwasprehybridized at42°C for 18 h and then hybridized with a random-primed or nick-translated 32P-labeled cDNAprobe(5 x105

cpm/ml)

at42°Cfor 24 h. After being washed, the hybridized filter was autoradiographed overnight at -70°C. The same blot was stripped of the labeled probe by incubation at 90°C in TE buffer (10 mM Tris-HCl [pH 7.5], 1 mM EDTA) for 15 min. The blot was then hybridized with the 5'-end _y-32P-labeled antisense oli-gonucleotides or a-32P-labeled cDNA probe for the yeast 28S rRNA gene. For the study of mRNA synthesis in the presenceof drugs, IE RNAswerepreparedfrom cyclohex-imide (50

,ug/ml)-treated

HEL cells harvested at 10 h p.i. Early RNAswereprepared from phosphonoacetic acid (100

,ug/ml)-treatedHEL cells harvestedat24 hp.i. Late RNAs wereharvestedat 60 hp.i.without drug treatment.

DNA sequencing. cDNA and genomic DNA sequencings were performedby the dideoxy chain termination method (19), with the Sequenase kit from U.S. Biochemical Corp.

(Cleveland, Ohio)andSequentide(NEG-034;NewEngland

Nuclear, Boston, Mass.). Single- and double-stranded tem-plates were obtained from appropriate subclonings in M13mpl8 and mpl9. The enzymes used for the subcloning include PvuII, RsaI, PstI, andXhoI. All sequences were obtained andconfirmed bysequencing data taken from both strands. Eight synthetic oligonucleotideswere used as

se-quencing primers to resolve sequence gaps and to confirm sequenceambiguities.ORFanalysis was carried out with the DNA Inspector II (Textco, West Lebanon, N.H.). Two-dimensional protein structures (4) and hydrophobicity and hydrophilicity profiles of the deduced peptide were deter-minedbyusingasoftware package provided by the Genetics

Computer Group (UniversityofWisconsin, Madison). Primerextension.Primer extension experiments were per-formed by the method described by Kingston (11), with minor modification. Four 30-bp oligonucleotides, including J. VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

(3)

STRUCTURALLY TRICISTRONIC GENE OF HCMV 2045

primer 517 (5'-ACACAAGCGAGCGAGTGGGGCACGGT GACG-3') and primer 520 (5'-CCACTCGAGAGCCTCCAT

GCGGGAGAGCAG-3'), both complementary to the 2.85-kb message, and primer 130 (5'-TGTTCGGAAACCGAGGCG GTGTCGCCATGC-3') and primer 136 (5'-TCGGCCACC

AGCGCGTGGCTGCGATGGAGC-3'), both complemen-tary to the 1.1-kb message, were synthesized on an Applied Biosystems synthesizer and end labeled with

[,y-32P]ATP

by using T4polynucleotide kinase. Seven micrograms of total RNA and 104 cpm of labeled primers were mixed for 5min at 80°C and then precipitated with alcohol. Pellets were dis-solved in 20,1 of lx hybridization buffer [80% formamide, 40 mM piperazine-N,N'-bis(2-ethanesulfonic acid) (PIPES, pH6.4), 400 mM NaCl, 1 mM EDTA] and allowed to anneal at30°C overnight. After ethanol precipitation, reverse tran-scriptase extension reactions were carried out at

42°C

for 90 min. These were done in a final volume of 25,u containing 18 U of avian myeloblastosis virus reverse transcriptase, 50 mMTris-HCl (pH 8.0), 5 mMMgCl2,5 mM dithiothreitol, 50 mM KCI, 50 U of RNasin, and 50

,uM

each of the four

deoxynucleosidetriphosphates. After RNasedigestion, phe-nol extraction, and ethanol precipitation, the pellets were dissolved in 4

RI

of TE and 4

RI

of gel loading buffer, heated at 80°C for 3 min, and loaded on a 6% polyacrylamide sequencing gel containing 7 M urea. The extension products wereanalyzed by comparison with genomic DNA sequenc-ing reactions, which were run in parallel. The control genomic DNA was derived either from a 1,008-bp (bp 3 to

1010)NcoI-PstI fragment upstream of the 2.85-kb mRNA or from a 602-bp (bp 1722 to 2333) PvuII-PvuII fragment upstream of the 1.1-kb mRNA.

S1 nuclease mapping. The four oligonucleotides used in primerextension were also used inS1 nuclease mapping (1). Primer 517 and primer 520 were used for the 5'-end mapping of the2.85-kb message. Primer 130 and primer 136 were used forthe 5'-end mapping of the 1.1-kb messages. The proce-dure used was as follows: 5'-end-labeled oligonucleotides wereannealed tosingle-stranded templates derived from the

correspondingNcoI-PstI region (for the 2.85-kb message) or PvuII-PvuII region (for the 1.1-kb mRNA), extended with Klenow enzyme, andrestrictedto generate uniform restric-tion fragments. A polyacrylamide gel electrophoresis-puri-fied probe (2x104 cpm) was annealed to 10,ugof total RNA in 30

RI

ofhybridization buffer (80% deionized formamide, 0.4 M NaCl, 10 mM PIPES [pH 6.5], 1 mM EDTA). This

annealingmixture was placed in a55°Cwater bath for 5min

and then incubated at 30°C for 16 h. To the annealed mixture, 200jilofS1buffer (30 mM sodium citrate [pH 4.5],

0.25mMNaCl, 1mMZnSO4, 5% glycerol) and 100 U of S1 nuclease wereadded. The mixture was incubated at

16°C

for 1 h. TheS1 nuclease-resistant fragments were precipitated with ethanol, denatured in 4 jI of formamide sample buffer

(90% formamide, 0.1% bromophenol blue, 0.1% xylene

cyanol), and electrophoresed in a 6% polyacrylamide

se-quencinggel with 7 M urea. The nuclease-resistant products were analyzedbycomparison with genomic DNA sequenc-ing reactions run in parallel.

Overexpressionof

US18,

US19,

and

Us20

in the pGEX-GST system. The GST gene fusion system was used to express viral DNAsubclones as fusion proteins (21). To construct in-frame clones, DNA fragments were inserted into the BamHIandEcoRIsites of one of three expression plasmids,

pGEX-1N, -2T, or -3X (21), with thereadingframematched in the sense orientation to yield recombinant plasmids

pGEX-lN-US18, pGEX-lN-US19, and pGEX-3X-US20.

Overnight cultures ofplasmid-transformed E. coli JM107

were diluted 1:10 with fresh medium and grown at

37°C

for 1.5 h. They were subsequently induced with isopropyl-3-D-thiogalactopyranoside (IPTG) at a final concentration of 0.1 mM. After 1 to 6 h of growth, cells were pelleted and lysed in 0.25 volume of SDS sample buffer (0.006 M

Tris-HCl

[pH 6.8], 4% SDS, 40% glycerol, 3% dithiothreitol, 0.005% bromophenol blue). The samples were boiled and subjected to 12% polyacrylamide gel electrophoresis.

The majority of the US18-GST,Us19-GST, and

US20-GST

fusion proteins were expressed as insoluble inclusion bodies. The inclusion bodies were prepared as described by Harlow and Lane (7) with modification. In brief, the induced plas-mid-transformed E. coli

JM107

cultures

(pGEX-1N-US18,

pGEX-lN-US19,

and

pGEX-3X-US20,

respectively) were

pelleted by centrifugation at 8,000 x g for 5

min.

The cell pellets were washed twice and resuspended at 1/10 their original volume in a washing buffer containing 100 mM NaCl, 1 mM EDTA, and 50 mM

Tris-HCl

(pH 8.0). Ly-sozyme was added to the suspension at a concentration of 1

mg/ml, and the suspension was incubated at room tempera-ture for 20min. The lysate was then sonicated for three 30-s cycles at full power with a Branson Sonifier (model 200 cell disrupter) equipped with a microtip and centrifuged at 5,000 x g for 10

min.

The

supernatant

containing soluble fusion protein was further purified on a glutathione-Sepharose column. The pellets containing inclusion bodies were washed with an ice-cold washing buffer (100 mM NaCl, 1 mM EDTA, 50 mM

Tris-HCl

[pH 8.0]) with 0.1% SDS, then with washing buffer containing 1% Nonidet P-40, and finally with washing buffer alone. The inclusion bodies were then further purified by 10 to 50% sucrose gradient centrifugation. Fractions containing inclusion bodies were pelleted and resuspended in phosphate-buffered saline (PBS), analyzed for purity, and used as antigens for antibody production in rabbits.

Antibody preparation. Antisera against the US18-GST,

Us19-GST, and

US20-GST

fusion proteins were made by individually immunizing New Zealand White rabbits. Inclu-sions with fusion protein aggregates were denatured in the presence of 7 M guanidine

HCI

solution and dialyzed against PBS with several changes of buffer solution to remove the guanidine

HCl.

A mixture of 100 jig of the fusion proteins with an equal volume of Freund's complete adjuvant (Cal-biochem Corp.) was injected into the rabbits subdermally at multiple sites near the hind legs. Booster injections with fusion proteins in an equal volume of Freund's incomplete adjuvant (Calbiochem Corp.) were given intradermally to the rabbits three times at three-week intervals. Ten days after the final injection, rabbits were bled, and the sera were tested for the presence of HCMV-specific antibodies by Western blot analysis and immunofluorescence staining.

Western blot and indirect immunofluorescence

analysis.

HEL cells were infected with

HCMV

as described above. At the indicated times, cells were washed twice with ice-cold PBS and then lysed with 2x SDS sample buffer (0.125 mM Tris [pH 6.8], 20% glycerol, 0.2% SDS, 0.28 M 3-mercap-toethanol, 10 jig of bromophenol blue). The lysates were sheared by passage through a 20-gauge needle four times and then boiled for 5 min. After electrophoresis, proteins were transferred to nitrocellulose paper in transfer buffer (20 mM

Tris-HCl [pH

8.0],

150 mM glycine, 20% methanol). The blot was washed three times with blocking buffer (2% dry milk in PBS) at room temperature for 10

min

each and then probed with antiserum against the fusion protein (diluted 1:100 in blocking buffer) at room temperature for 1 h. The antiserum was removed, and the blot was rinsed three times with a

VOL.67, 1993

on November 9, 2019 by guest

http://jvi.asm.org/

(4)

2046 GUO HUANG

A. iCcV 0.0

gsne

IL

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

I I I I I I I I I.

B EoRI LT M C VOKUT E J N G R A S B LZXH I U/S F D WST

map II I 1111I I I I IIIII1 I I _I_

C Transcript

Xho lndE Hind RO

ago

lindl

.i I *I I I I I I I

E-OM

n i I

_

5

Transcdfte

D. pbe

Hhid ag.

a-.A, I|I

T..AAA

4-=M

1.1Kb mRNA

E cDNAmd 5 en 3'..A -..

2.2KbCDNA

V...Am

10 Kb

g.o

5

primer 517 & 520

prime 130& 136

Is2bP !

SIbp

1.1 KbCDNA Om

F. Arrangementofputative geneproducts USi8

3'..AAA W

822 bp

5 end CAP ste

i CAPaft

USi 9 IH 720bp

US20

14 1026bp

.1

s.

HWLF5 HWLF 4 HWLf 3

FIG. 1. PhysicalmapofHCMVTowne strain DNA and relativelocations of theEcoRI Dfragment, cDNAs,and varioustranscriptional

products derived fromthis tricistronic gene.(A) Fractionalmapof theHCMVDNA genome.(B)Schematicrepresentationof theHCMV genome,with the restrictionmapfor EcoRI.(C)EcoRIDfragmentof theHCMVgenome,designatedpHD4(5).Thetranscriptionorientation ofthe 2.3-kbcDNAisindicatedbyanarrow,and themagnifiedmapis shown underneath.(D) Transcripts andrelative locations ofDNA primers used forprimer extension and S1mapping.(E)Schematicdiagram showinglocationsof the2.85-kb and 1.1-kbmRNAsandcDNAs and their locationsrelativetotranscription initiation sites.A628-bpsequencewasmissinginthe 2.3-kbcDNAandan81-bpsequencewas

missingin the1.1-kb cDNAcompared withtheirmRNAs.(F)Thelocation and arrangement ofputativegeneproducts.Datawereobtained from theexperiments showninFig.2through8.

blocking buffer. Bound antigen-antibody

complexes

were detected by incubating the filter in buffer solution with

protein A-peroxidase-labeled secondary antibody (Boehr-inger Mannheim Co.) at a 1:500 dilution for 2 h. After

incubation, the filterwaswashedasabove and reactedwith substrate(10mgof diaminobenzidine in 100 ml ofPBSwith 10 ,ul of30% H202)for 30 min.

For theindirectimmunofluorescencetest,antibodies were preabsorbed with HEL cellular lysates at 37°C for 1 h. HCMV-infected HEL cells grown on eight-well chamber slides (Nunc, Naperville, Ill.) were harvested at the indi-catedtimes p.i. The cells were washed with PBS three times andfixed with ice-chilled methanol-acetone (1:1) at -20°C for 15 min. The fixed cells were then blocked with 20%

normalgoat serum(JacksonImmuno Research Laboratory

Inc.)inPBS for 30 minatroomtemperature.Cells were then reacted withdiluted rabbit antiserum (1:50 in PBS) for 1 h. After PBS washes, biotin-labeled goat anti-rabbit

immuno-globulinGsecondary antibodies (VectorLaboratories Inc.,

Burlingame, Calif.) at a 1:200 dilution were added to the

cells and incubatedat roomtemperaturefor 1 h.AfteraPBS

wash,fluorescein-conjugated avidinD(VectorLaboratories

Inc.)was added at a concentration of 12.5 ,ug/100ml. The slidewasmounted with50%glycerolinPBS and examined underafluoresent microscope.

Nucleotide sequence accession number. The GenBank ac-cession number for the sequencereportedinthis

communi-cationis L04998.

RESULTS

cDNA cloning. In our first attempt to isolate EcoRI D

fragment-related sequences, a 2.3-kb cDNA was isolated from our Xgtll cDNA library made from HCMV-infected fibroblasts. This 2.3-kb cDNAwasmappedin the EcoRI D

fragment.Themappingwasdoneby Southern hybridization and further confirmed by DNA sequence analysis at both ends of the 2.3-kb cDNA, as shown below (Fig. 1). In subsequent screenings with the 2.3-kb cDNA clone as the

probe,80positivecloneswereisolated from 4 x 104

recom-I

on November 9, 2019 by guest

http://jvi.asm.org/

[image:4.612.157.472.72.427.2]

(5)

A.

B. C

Mock Cyto Total Nuci Mock Cyto Total Nuci Mock Cyto Nuci Total

MW(Kb)

9.49-_

7.486

4.40 -_

2.37 -- _2.85

[image:5.612.100.529.77.285.2]

1.8 1.35 -_

FIG. 2. Transcription analysis with the 2.3-kb cDNA and strand-specific oligonucleotide probes. Total RNA (8pLgperwell) was prepared from mock-and HCMV-infected HEL cells. Cytoplasmic, nuclear, and total RNAs were purified from HEL cells at 60 h p.i.Cytoplasmic and nuclear RNAs were furtherfractionated byoligo(dT)-cellulose column chromatography to obtainpoly(A)'mRNA. Sizes are indicated on the left, andarrowsonthe right indicate themessages detected and their sizes (in kilobases). (A) Probed with astrand-specific probe (primer 136) close tothe 3' end. Twomessages, 2.85 and 1.1 kb, were observed. (B) Probed with astrand-specific probe (primer 520) close to the 5' end of the 2.85-kbmRNA.Thisprobedetected only one message of 2.85 kb (see Fig. 1forrelative locations of oligonucleotide probes used in hybridization studies). (C) Probed with labeled 2.3-kb cDNA. This probe detected 2.8-kb, 1.1-kb, and a minor 1.8-kb message.

binant phages. Twenty positive clones were analyzed by restriction enzyme mapping and DNA sequencing. The clones were categorized into three groups, with a cDNA insert of 2.3, 1.8, and 1.1 kb, respectively. Three cDNA clones were further cloned into M13mpl8 for 3'-end and 5'-end analysis. Theresults of 3'-end analysis reveal that the 2.3-kb, 1.8-kb, and 1.1-kb cDNAs have an identical poly-adenylation signal (AATAAA) and a 14-bp downstream

sequence(GCTTT1TT1-1CACGCTl) and are3' coterminal.

Transcription pattern. To analyze the transcripts corre-spondingtothe2.3-kbcDNA, RNA samples isolatedat60 h after infection were probed with the nick-translated and random-primed labeled 2.3-kb cDNA fragment. Two major

groupsofmRNAs of2.85and 1.1kbweredetected (Fig. 2C).

Thedetection ofaminorband of 1.8kbwas notconsistent andvaried from batchto batchofRNA preparations. Hy-bridization withanantisense oligonucleotide probe (primer 520)derived fromthe 5'endof the 2.3-kbcDNA detected the 2.85-kb message but not the 1.1-kb transcript (Fig. 2B). When an antisense oligonucleotide probe (primer 136) de-rived fromthe 3' end of the 2.3-kb cDNAwasusedasthe probe, both the 2.85-kb and the 1.1-kb messages were

detected (Fig. 2A). This result suggests that the 1.1-kb transcript wastranscribed within the 3'-end portion of the 2.85-kbmessage. The DNA sequencebetween thepoly(A)

signal sequence and the 5' end of the 1.1-kb cDNAwas

equivalent tothe length detected in the 1.1-kb RNA tran-script. These data suggest that the 2.85-kb and 1.1-kb mRNAs are 3' coterminal and that the 1.1-kb transcript is unspliced. These3'-coterminal transcriptswerefurther con-firmedby3'-endsequencingofthe cDNAaswellasbythe methodoftherapid amplificationof thecDNA ends(RACE) (datanotshown).

5'-endmapping by primerextensionand Si nuclease

map-ping.Twoantisenseoligonucleotides (primer517 andprimer 520), complementaryto twoareasclosetothe5' end of the

2.85-kb mRNA, were used in the primer extension studies

(datanot shown). Both of these primers showed the same

major transcription initiation site right at the translation

initiation codon of the

Us20

ORF, even though these two

primers are 80 bp apart. These two oligonucleotides were also used in

Si

nuclease mapping. Themajor transcription initiation site for the 2.85-kb mRNA was similar to the

primer extensionsite (Fig. 3A).However,whenprimer 517 was used for S1 nuclease mapping, a minor band also

appeared about243bpdownstreamfrom the major initiation site indicated above, which is approximately 26 bp

down-streamfromaputativeTATAboxandapproximately21 bp

upstreamofapotential second initiation codon for the

Us20

ORF(datanotshown).

Anothersetofantisenseoligonucleotides

(primer

130and primer 136), complementary to the 5'-end region of the 1.1-kbmRNA, wereusedin the primer extensionstudiesto detect thetranscription initiation site of the 1.1-kb mRNA. This set ofprimers picked up four transcription initiation

sites with equal intensity within a 7-base range

(data

not

shown). S1 nuclease mapping withthesameoligonucleotides

alsodetectedamajortranscriptioninitiationsiteinthesame

region (Fig. 3B). This sitewas 33bpupstreamof the

Us18

translation start site. The 5' end of this transcript has a

TATA-likebox, GATAAGA

(bp

2115to2121), 22to 28

bp

upstream of the transcription initiation site. A CAAT-like

box, CGTCAATCA (bp 2085 to 2092), was 50 to 58 bp

upstream of thetranscription initiation site. The

transcrip-tioninitiationsitewas 33bpupstream of the

Us18

transla-tioninitiationcodon. Theschematicdiagramsin

Fig.

1E and F outline the transcription initiation sites found and their relative locations in the cDNAs. The detailsareindicated in

Fig. 6.

Attempts to map the 5' end of the minor 1.8-kb mRNA were not successful when

primer

156

(5'-GTAGCAGCAG

GAACA-3')

wasused for

primer

extensionand

S1

nuclease

VOL. 67,1993

on November 9, 2019 by guest

http://jvi.asm.org/

(6)

2048 GUO AND HUANG

L T G C A

i.. 5.1

G T

T G C A E

B

T T

c A T

c

G A c G

c

T

* A

_ _

_-I

- _A

- a

___w

so=

wrwm

_ _ mo

4

__ _ _w

c , i_

T I

T A G A

C J:

C I

c

G |

A c

- a

-p

-u

-a

"o

G-FIG. 3. Si nucleasemappingof the2.85-kb and 1.1-kbmessages.The5'-end-labeledoligonucleotides (primer520 for the2.85-kbmRNA, and primer 136 for the1.1-kbmRNA)wereannealedtocorrespondingsensesingle-strandedDNAtemplatesfor thesynthesisof32P-labeled antisenseprobes.Thepreparationof radioactivesingle-strandedDNAfragmentsand theprocedureforS1mappingaredescribed in Materials

andMethods. Thedideoxynucleotides (T, G, C,andA)used in the reactions areshown above the lanes.The DNAsequencingdata from

parallel experimentsareshowntotheleft of eachpanel;possible transcriptioninitiation sitesaremarked with asterisks(*).(A) Si mapping

ofthe 2.85-kb mRNA. Late RNA(L)isolatedat60 hp.i.wasused in thisanalysis. (B)Si mappingofthe 1.1-kb mRNA. RNAs isolatedat

24(E) and 60 (L)hp.i.wereused in thisstudy.Thetranscriptioninitiation site is indicatedbyan arrowand markedbyanasterisk(*)in the

sequence.

mapping, perhapsbecause of the low transcription level of thismessage.

3'-endmapping bycDNAsequencingand RACE. Inorder tofurtherverifythat these mRNAswere3'-endcoterminal, two different methods were used: sequencing different batches of cDNA, and RACE. Twenty positive cDNA clones were sequenced at both ends, and the sequence

results indicated that these threegroupsof cDNA(2.85, 1.8, and 1.1kb)usethesamepolyadenylation signal (AATAAA), which is 15 bp upstream of the poly(A) tail. On the other hand, these three cDNAs have differentlengths ofpoly(A) tail. The results of the RACE experiment revealed asingle band of 290 bp, the predicted size. Subsequent sequence

analyses of this RACE-amplified DNA fragment further verified the results (data not shown). Both the cDNA

se-quence and RACE amplification provide supporting data for 3'-end cotermination of the major 2.85-kb and 1.1-kb mRNAs and theminor 1.8-kb mRNA.

Analysisoftranscriptionpatterns.TotalRNAscollectedat different time points (0, 4, 8, 12, 24, 48, 60, and 72 h p.i.) wereelectrophoresed througha1.2%agarosegelcontaining 6%formaldehyde and analyzed byNorthernhybridization. Thetranscriptionpatternswerederived from theuseoftwo

antisenseoligonucleotide probes,onefrom the 5'end of the 2.85-kb mRNA and another from the 3' end of both

tran-scripts. The5'-end probe detected only the 2.85-kb mRNA (Fig. 4),while the3'-end probedetected both the2.85-kb and 1.1-kbmessages(Fig. 4B). RNAlevelswerequantitatedby densitometerscanningandnormalizedtothat of the internal control, rRNA. Transcription of the 2.85-kb mRNA first appeared at 12hp.i.,reachedamaximumlevelat48 h p.i.,

and thengraduallydecreased. The 1.1-kb mRNAappeared between 24 and 48 h p.i. and remained ata constant level until 72 hp.i. (Fig. 4C). Theincrease intranscriptionat96h p.i.wasprobablyduetovirusreinfection.

Transcription patternin thepresenceofdrugs. To further characterize these two mRNAs, the transcription patterns were studied in the presence of protein and viral DNA synthesisinhibitors. In thepresenceof theprotein synthesis inhibitor cycloheximide (50 ,ug/ml), both the 2.85-kb and 1.1-kbtranscriptsbecameundetectable (Fig. 5). These data suggested that the 2.85-kb and 1.1-kb messages might re-quire HCMV IE gene products for the initiation of their transcription. Therefore, we conclude that these two

mes-sageswerenotIE transcripts.

In the presence of phosphonoacetic acid (100 ,ug/ml), a viral DNAsynthesis inhibitor, the 2.85-kbmRNA accumu-lated to a level comparable to thatin the control (without drug) at 60 h p.i. The transcriptional level of the 1.1-kb

messagein thepresenceofphosphonoaceticacid, however,

decreased to less than one-fifth ofcontrol levels (Fig. 5). These data suggest that the 2.85-kb mRNA is actually an early, or a-type, message ofHCMV, which didnotrequire viralDNAreplicationforexpression. The 1.1-kb mRNAwas expressedat latetimesduring infection(24to48 h p.i.) and requiredviral DNAreplication for maximumexpression; it is clearlya late,or -y-type, message. Theresults show that thesetwotranscriptswereexpressedatdifferent times after infection andwere regulated by differentmechanisms.

Computer analysis of DNA sequence. A 3,158-bp DNA sequence (Fig. 6)was assembled from the sequencing data determined from the analysis of both strands of various

A

5t

C G A G T A T *G A G A T T A T A T C C

3'

., -.

A

.;A.4.;

on November 9, 2019 by guest

http://jvi.asm.org/

[image:6.612.147.484.79.322.2]

(7)

STRUCflURALLY TRICISTRONIC GENE OF HCMV 2049

MW._(Kb;f mock 4 8 12 24 48 60 72 96

A *

4o

_ 0 *2.85 Kb

2 37 *

M W (Kb) 9.49 -_

7.46 -_

Mock CH PAA LATE

4.40

-135 4

2.37

--4.40 *

1.35-_

B

2.85Kb

237 *

1.35 *

60

I6*

1.10Kb

C

[image:7.612.341.525.76.221.2]

440 4

FIG. 4. Transcriptionpatternanalysis ofthe 2.85-kb and 1.1-kb mRNAsby Northern blothybridization. TotalRNAwasprepared

frommock- and HCMV-infected HEL cellsatvarious times(0, 4, 8, 12, 24, 48, 60,72,and96hp.i.)asindicated aboveeach lane. Eight

microgramsoftotalRNAwassubjectedtoelectrophoresis through

a1.2%agarosegel containing 6% formaldehydeandthentransferred

tonitrocellulosepaper. (A) The resulting filterwasprobed withan

antisenseoligonucleotide primerlocatedatthe 5' endof the2.85-kb mRNA(seeFig.1D, primer 520). (B)The filterwasreprobed withan

antisenseoligonucleotide primer (primer 136)locatedatthe 3'end of both mRNAs. (C) The blot was probed with yeast rRNA as an

internal control. The arrows on the right indicate the messages

detected, and the arrows onthe left indicate thepositions of size

markers(in kilobases).

restricted genomic or cDNA fragments with an M13mpl8 subcloning system and primer extension approach. This DNAsequence includes a 2,290-bp cDNA fragment and a 1,008-bp genomic regionupstreamof the5' end of the 2.3-kb cDNA. ThesetwoDNAfragmentsoverlaptailtohead,with 140bpatthe3' end of the1,008-bpupstreamgenomic region and at the 5' end of the 2,290-bp cDNA sequence. ORF analysesof thesequencesshown inFig.1FandFig.6were

donewith the DNA Inspector II program. Three

indepen-dentnonoverlappingORFs of Towne strainHCMV, equiv-alent to

Us18, Us19,

and

Us20

of AD169 strain HCMV, werederived. Atypical polyadenylation signal (AATAAA, bp3138to3143)waslocated15bpupstream of thecommon

poly(A) tail. All of these ORFs have a typical consensus

Kozaksequence(12)around the translation initiation codon. Sequence comparison with AD169. The sequences of the Towne and AD169 strains of HCMV revealed 99.3%

homol-ogy.The Towne strainisanHCMVstrainisolated from the urine ofacongenitallyinfectedbaby. The AD169 strain isa

prototypeofHCMV isolated from adrenalglandsalmost 25

years agothat has been subculturedinvitro formore than 280passages(17).Thesequencecomparisonbetween these twostrains revealed 20 base mismatches(Fig. 6). No inser-tions or deletions were found. Nine of the 20 mismatched bases are locatedwithin noncoding regions. Ten of the 11

othersingle-basemismatches do not resultin anyaminoacid

sequencealteration, andtheothermismatchchangesamino acid residue 238 of

Us20

from alanine in AD169 to valinein

FIG. 5. Transcription inthepresenceofprotein and viralDNA synthesisinhibitors.Eight microgramsof total RNAfrommock- and HCMV-infectedcellstreated withcycloheximide (CH;50 p.g/ml)for 10horphosphonoacetic acid (PAA;100 ±g/ml)for 60 horuntreated

cellsat60 hp.i.(LATE)wasappliedtoeach laneforNorthernblot RNAanalysis.Theresultingfilterwasprobed with the 2.3-kb cDNA

probe.Thearrows ontherightindicatemessagesdetected, and the

arrows onthe leftindicate the migration of sizemarkers.

the Towne strain. The results of this experiment show that the products of this tricistronic gene are very well

con-served.

Two-fdimensional protein structure analysis. Three ORFs

(US18, Us19,

and

Us20)

were determined from DNA

se-quenceanalysis. Two-dimensional structure analysis of the putative

Us18

polypeptide reveals seven hydrophobic

seg-ments. Each ismorethan20 aminoacidslongandmayspan

the cell membrane. One strong antigenic site was found withinoneof thetwopotential glycosylation sites, where the hydrophilic moiety is also obvious. The N-terminal end of this peptide is hydrophilic, whereas the C-terminal end is morehydrophobic.ThisORF has40leucineresiduesamong

274 aminoacids. Theputative

Us19

polypeptidehasonly six hydrophobicsegments,but oneof these isalongstretchof 50hydrophobicamino acids. Nopotentialglycosylationsites werefound in the

US19

polypeptide,buttwohighlyantigenic clusterswith hydrophilicmoieties were found between the first and the second and between the second and third hydrophobic domains from the N-terminal side of this

pep-tide. Both the N- and C-terminal ends of this peptide are hydrophobic. The putative peptide contains 52 leucines

among240 amino acid residues. Two-dimensional structure analysisof

Us20

revealedsevenhydrophobicsegments,two potential glycosylation sites, and twohighlyhydrophilicas wellashighly antigenicmoieties located around amino acid residues 50 and 250. Theputative

Us20

peptidecontains 47 leucinesamong342 amino acid residues. Thehydrophobicity profilesof thesepeptidesshow characteristics of membrane-associated proteins: long stretches of hydrophobic amino acids connected by alength of random coil or hydrophilic amino acids.

Western blot analysis and indirect immunofluorescence tests.Antibodiestothebacteriallyexpressedfusionproteins were able to reactwithvirus-specific products in Western blotanalyses.The results obtained with HCMV-infectedcell

lysates showed that

anti-US18-GST

antibodies detected a 36-kDavirus-specificband(Fig. 7A). Thisprotein appeared at 34hp.i. andcoincided with the latetranscriptionof the

1.1-kbmRNA,a-y-class transcript.The

Us19-GST

antibody detecteda32-kDavirus-specific peptidewhichappearedata

_l _ - 2.85 Kb

_1.1 Kb

VOL.67,1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:7.612.64.291.78.293.2]

(8)

2050 GUO AND HUANG

NcoI IR1

CA*IUCCA TA A GIN1II=O=CTACGTA7TrGCC=

i2 IR1

G C C

IR-3_1 DR1,_ r

CGCCG 1T rVCT=CGAGGCGGTA,AMAATGCP GTGAATC1 AGOCGGIGCG7CACGIGCATOTG7GGTACGCXGACAClT%CPGACATGA

IR3

CAP2.85 K0rtA

SnsI IR4_ -6IR4

MAGAGTTTAl.Q:C=TCGGCGACA7GACClt T3TlACCACITG7GTACCTACA ATGCOCI

R V I S R A R S A C. T W T S C T S L S P C S T S C P

IR5 IR6 IR6

DR2 ,_DR2,_

G A a s T G T~~~~~~GA37Ga3GCGrACCA: ACGTGCa:XCACTI= P Q Q R R R P S S S P N R R V R G V T T S P C P T R

r minor DR3 e_ DR4 DR3 __ Xho

TrrGCGAGACCGTCGTCACCALAGGCGkGCTACGCOs iG=aCATG'A GGCT9og

r*2ndtransl. init.

L C R DR RHIM Q A Q E A N A L L L S R M E A L E W

WATCTAGCIT.GClAGCTIGGCOTIOGGAAGCG CI±CA AAAC OAACl CG I F Q L A F S P G L G S V F W L G F P Q N R N F C V TCTGCATGC ATCACG7ACACGTIGGGCAACGAACACCCTAGrAACGrCACGGroCC1I7ATCM¶O¶IGTI7GCCAAM

C M F I T Y T LN.A.TG N E H P S V L F I Y L L A N

r-__cDNA2.3 kb

.=CGArJCCCGrAGCGOCCAC

P S P A A P T

TATABOX

.~O~ AA G

S L V Y K R R ,TCAAAAAGTTCACCGrATG

F K K F T V W 3GAGAACTACPCCICTI

E N Y F F L .AGCCrGACGGCGGCCATCTI S L T A A I F

H

M~

IR5

ACI CTCArACGA0GGTCGC

L L R R R S L

GGrAGGT1XGCCGCAGGGC

V G A P Q R L GCTGCGCGIGTACXCATCT L R V Y A I F

CACCGlrxCGTGOZCATCG

T V L V P I V CCAGA7G7GCTCTAAAGCC

A

Q M C S E S R

GCGTACrAGTAG= AI G T C TGIACC C I G

, C G T ~~~~~~~~~~~GACCAC

A V L V G S Y V M T L A L F I S F T G L A F L G G R D R R R W K C I S C V Y V V M

PstI

TGTGC=GIrI CGTrCTCTGCAGCCPGCCAICT CTElIAAGATAATAGT

L L S F L T L A L L S D A D W L Q K I V V T L C A F S I S F F L G I L A Y D S L

TCATGGA7TCII'TC7lCCAOCTAACCAATGATCCTGCCGICTGCACCIGACGCA C CACCGTGATA

M V I F F C P P N Q C I R H A V C L Y L D S M A I F L T L L L M L S G P R W I S

TATA BOX

GTCTTGGCGCGAr,C711CGkCAACOSGAC'TlrAArGCCACGr CACGACGGA AGTCCTARGATCCGCCGr=CAGCA0CGTaCA=CMCAGC

C T

L S D G V P L D L T A A S T T G K S * M L A

CATG=YxTCCCGCTAGAATGGAC0GTCGAGGAGGT=CCCrCACCTAGAAGA'GGGTATGGCSAGOSCTGC3CCCGTCC7K)TCCCCA(TTACOGCCACOCGTCGCAGC

H V V P L E W T V E E V V P Y L E R L A V W L R A S V L V A F Q L T A T V A L S

GTr.CTCAGCC.ITG~TGCCOCACCGr.TG..C3GCTGArCGCATGGGGGGACGGGA0ACACCCCGCGC ACCACCAOTGATTGCCA GGCIrCCTSTGC V L S W W L M P P P V A E L C E R G R D D D P P P L S H L S L V V P V G C L F L

CT,GCTACGCGGsICCGTCGkTCGAOC GTTGlUCCCGGmAACTCOCTCITGTIGCrACAGCT(rCGCA0GCGCTAGCCITCAC IGCrC-7G'AGCCA'MCG:AGGCC

L L R G P S I D R C P R K L P L L L A Y C L P H A L A F L T L L M C Q P S P Q A

DR5,._ DR5 ,_

IIIIIrGGCGC(33CloGC(rlA3TGAlCTAGCICGGCCllTlCGC' lleTrCqYGACC.,-GCGTCICGCCG(CIr-GCTACCGTC(GrICYVCTG

F V G A A L L A L A V D L S C L G A S L L G C D P G A S L R R L W L P S V L S L Spl/IR7 DR6 ,_ DR6 ,

CTCTGCGCTACGGG=GCGCTCGGSCCGC G:CCCCTCGTrIr=GTTACA a3CGACGkCGACGIrACMr.AC(TGT<TAATwACGA7CG7ATC

Spl/IR7

L C A T A L G L W L L R A A A P F F L G L H A T T L L T V T L M L I H D L S L I

120

240

360

480

600

720

840

960

1080

1200

1320

1440

1560

1680

1800

1920

Towne Genomic

AD169Genomic

TowneGenomic AD169 Genomic TownePeptide

Towne Genomic AD1 69 Genomic TownePeptide

Towne Genomic

AD169 Genomic TownePeptide

Towne Genomic AD169Genomic TownePeptide

AD169Peptide

Towne Genomic AD1 69 Genomic TownePeptide

AD169Peptide

Towne Genomic AD169 Genomic TownePeptide

AD169Peptide

TowneGenomic Towne cDNA AD169 Genomic TownePeptide

AD169Peptide

TowneGenomic Towne cDNA AD169 Genomic TownePeptide

AD169Peptide

Towne cDNA AD169 Genomic TownePeptide

AD169Peptide

TownecDNA

AD169 Genomic

TownePeptide

AD169Peptide

TownecDNA AD1 69 Genomic TownePeptide

AD1 69Peptide

TownecDNA AD169Genomic TownePeptide

AD169Peptide

Towne cDNA AD169 Genomic

TownePeptide AD169Peptide

Towne cDNA

AD1 69Genomic

TownePeptide

AD169Peptide

Towne cDNA AD169 Genomic

TownePeptide

AD169Peptide

OG ;c

SG :G

IY LQ

on November 9, 2019 by guest

http://jvi.asm.org/

(9)

DR7 ,_ DR7 ,_

2040

T C Q S S F P E S F Q P S L R L Y V E N V A L F I G M Y H L L R L W L W S P *

F CAP1.1 KbufA

CAATBOX TATA BOX

2160

r-11CDNA 1. 1 kb

M G D T A S V S E H H E S P T V T I V P L H R S H A L V A EQ Q L F Q W

w T CZI~ZZ~CGGGACAGTGCG

T

L K R F K L L M E V Y H G L V W Q L A C T L T V C L L A W L A F P DV Q G Q C A

A A T

N G I V P A L S S I V P V S T L A M L R G F A E F R P H T T N F A H L T V A C L

~~ ~~CN3CGGTIlrllACC¶ACIQG

L I N T G I T V C T G F C G E R R V I G L S F A L V M V F F V L C S G L T Y L A

CCGGCACAAZ0ACAACTGTCATCACA3CGGAC

C C

G N N_L_P R W K V I GI G Y G W S V I V F Y L L L Y F S P V L W V S K I Y S G L

Y V L V V T A A S A V L I Y E T L D L I Y Q R G T L S K N S V C V S V V L Y T I

7 ATCGGI'rC ATAMACA CTCPr ATAAA

A G

V M S L L N M S V A I F S G H V W V Q Q Y A E K H G G R I D G V S L L S L L *

AAGAT1 GGAITGC IC,CCCI=GAAAGGTC

C G A T T

POLY ASIGNAL

CGTT1ITmATI'CAAC6T _ _ _CGTkAskssskhshksskkhksAkkkk_kkhs

2280

2400

2520

2640

2760

2880

3000

3120

3209

FIG. 6. Nucleotidesequenceof thestructurally tricistronicgeneencodingthe2.85-kbmessageand theputative polypeptidesUs18,Us19,

andUs20.The2,290-bp cDNA(2.3-kb cDNA)and1,008 bpofgenomicsequencelocatedupstreamofthiscDNAwerecombined andare

shown inthisfigurewithanoverlap of 140 bp. Sequence mismatches between the Towne andAD169 strainsarenoted intheAD169genomic

sequence.Thepolypeptidesequencesareindicated belowthe DNAsequence.Thetranscription initiationsitesidentifiedby primerextension andS1nucleasemappingareindicatedbysolid bentarrows.The minorcapsite forthe 2.85-kb mRNA is also indicated. Thebeginningsof

thetwocDNAs used in thisstudyarealso marked withopenbentarrows. Upstreamsequenceanalysisof the2.85-kb and 1.1-kb mRNAs revealedseveral indirectrepeats(IR)and directrepeats(DR),whicharenumberedsequentially.Solidhorizontalarrowsindicate the IR and

openhorizontalarrowsindicate the DRupstreamofUS18andUs20.Thepolyadenylation signalsequenceandpotential glycosylationsites

areunderlined. Thepossiblesecond translation initiation codon forUs20isindicatedbyabentarrow.

verylow levelat18 hp.i.andwasclearlydemonstratedat34 h p.i. (Fig. 7B). The

anti-Us20-GST

antibody detected a major 43-kDavirus-specific product at 18 hp.i. (Fig. 7C). This translation result also matched the 2.85-kb mRNA transcription data. A minor bandof36kDawas foundata

late time(34h)in virus-infectedlysates.This 36-kDaprotein mightresult from theuseof the second translation initiation codonatlater stages of virus infection.

Antibodies generated in a rabbit against the bacterially expressed fusion proteins

(Us18-, Us19-,

and US20-GST)

gave positive immunofluorescence staining in HCMV-in-fected cells. This reactionwas not found in mock-infected cells orat 12 h after HCMVinfection. The anti-Us18-GST

antibody detected positive immunofluorescence only at 48 and 72 h p.i. (Fig. 8). This result suggests that this

virus-specific productwastranslatedat48h,whichcoincided with thetranscriptiondata for the 1.1-kb mRNA. The

Us19-GST

antibodiesdetectedpositiveimmunofluoresenceat24, 48,and 72 hp.i.,with patterns similartothat with the

anti-Us20-GST

antibody (datanotshown).Thisresult indicates that theUs19

virus-specific products appear earlier than the

Us18

virus-specific products.The

US20-GST

antibodiesdetectedpositive immunofluorescent reactionsat24, 48,and72 hp.i.,likethe

Us19-GST

antibodies, but the reactionwasstronger than it

was with the

Us19-GST

antibodies at 24 h p.i. (data not shown). This result indicates that the Us20 virus-specific products appeared earlier than the Us18 and US19 virus-specific products. The early translation of the Us20 virus-specific products also matches the 2.85-kbmRNA transcrip-tion data.

VOL. 67, 1993

Towne cDNA AD169 Genomic TownePeptide AD169Peptide

Towne cDNA

Towne cDNA AD169Genomic TownePeptide

AD169Peptide

Towne cDNA AD169Genomic TownePeptide

AD169Peptide

Towne cDNA AD1 69 Genomic TownePeptide AD169Peptide

AD169Peptide

TownecDNA AD169Genomic TownePeptide

AD169Peptide

Towne cDNA

TownecDNA

ACC'=AGAGTrC7q'rO::.CAG'rrrICAGCC

on November 9, 2019 by guest

http://jvi.asm.org/

[image:9.612.69.555.69.474.2]

(10)

DISCUSSION

MW

Mock 5 18 34 48 72

Genomic repetition is a common characteristic of the humanherpesvirusgroup, but gene clusters inthe

Us

region are a unique characteristic in HCMV and herpes simplex virus (3, 24). Before the HCMV genome had been com-pletelysequenced, weoriginally speculatedthat the Towne strain EcoRI D fragment located within the

Us

region

encoded theHCMV counterparts ofherpessimplexvirusgD andgC. Inourfirst attemptto screenforgD-relatedcDNA, a2.3-kb cDNAwasisolated fromaXgtllcDNA

library.

This cDNA has been used as a tool to characterize the

corre-spondinggene. Tooursurprise, insteadofisolating gCorgD homologs, we identified the

US18, US19,

and

US20

tricis-tronic gene.

In the study of the transcription products related to the 2.3-kb cDNA, twomajormessages (a2.85-kb mRNA and a

1.1-kb mRNA) and a minor 1.8-kb mRNAwere identified.

Detection of the minor message by the 2.3-kb cDNA probe wasnotconsistent,varyingfromonepreparationtoanother. FromDNAsequencingand ORFanalysis,theminor 1.8-kb productmight bederivedfrom the secondORF,which does nothaveastrongclassicalpromoter-enhancer sequence and mightrequire special cellularorviral products for maximal expression. Theexpression of thisgroupof messagesmight

be under very tight control. The presence of the 1.8-kb cDNA reflects the existence of this weak message.

Mean-while, the detection of 2.85-kb mRNA suggested that the 2.3-kb cDNAisnot afull-length cDNA,since

approximately

628bp of the 5' upstream sequencewasmissing. Also, the 1.1-kb mRNAcould represent the 1.1-kb cDNA.

As determined by primer

extension,

the 2.85-kb mRNA has twotranscription initiationsites;oneisrightat the first translation initiation codon of the

US20

ORF, and the other is 243bp downstream from the first site. There isa typical

TATA sequence (TATAAGA) 21 bp upstream of the first transcription initiationsite, but no typical CAAT sequence wasobserved. Therewere twoGC boxes andadirect repeat located about 40to80bpupstreamof thistypicalTATAbox. On the otherhand, no typical TATA

[TATA(A/T)A(AJT)]

box or CAAT

[GG(C/T)CAATCT]

box upstream of the

second cap site was found. It is unclear why the major transcription initiationsite for the 2.85-kb mRNA islocated right at the first translation initiation codon of

Us20.

In

general, leader sequences are20to 100nucleotideslongon most vertebrate mRNAs except for mRNAs derived from proto-oncogenes, which have very long leader sequences

(12).Inaddition,the ATG codon isnotwellrecognized when it occurs close to the cap (12). Therefore, the translation regulation of the

Us20

ORF needs further investigation

before anyconclusionscan bemade.

Thetranscription initiation site of the 1.1-kb mRNAwas 33bp upstream of the translation initiation site.Intheregion upstream ofthistranscript, there is an atypical TATA box sequence(GATAAGA)andanatypical CAAT box sequence

(CGTCAATCA)

within 22to28bp and 50to58bp upstream

ofthe cap site, respectively. Promoter activity upstream of the 2.85-kb mRNA was detected during HCMV infection

(data not shown). The promoter activity of the region upstream of the 1.1-kb mRNAis still under investigation. Withregard to the

Us19

ORF, there are typical TATA and

GCboxes upstream of the translation initiation site and a minor 1.8-kb mRNA which could correspond to a transcript

initiatingnear these potentialtranscription start sites. How-ever, this message was not clearly demonstrated by

Si

mapping analysis.

46 K

A--*-36 kD

30 K

--.. 1,

21.5 k -.3m- "M,¶I:

46 K-- B

30 K

-1-21.5 K

--46 K -i C

--32 kD

-- - -. 43 kD

-K-36 kD

[image:10.612.329.553.79.284.2]

30 K -E

FIG. 7. Western blot analysis of HCMV-infected HEL cellswith

anti-Us18-GST, anti-Us19-GST, and anti-Us20-GST antibodies. DNAfragmentscorresponding to ORFsUs18, Us19,andUs20were isolatedby restriction digestion and then subcloned into the pGEX-GSTgene fusion system forexpression in E. coli(21). Thefusion proteins obtained were used to generate antibodies in rabbits for immunological analysis of these peptides. HCMV-infected HEL cells(2x 105cells) harvestedatthetimes indicatedabove the lanes (hoursp.i.)wereused for Westernblotanalysis.Arrows ontheright indicate virus-specific bands, and arrows on the left show the positionsofsize markers (inkilodaltons). Antibodiesderivedfrom the (A) Us18, (B) Us19, and (C) Us20 fusion proteins detected virus-specific bands of 36, 32, and 43 kDa,respectively.

Sense-strand oligonucleotides targeting either the 5' end orthe 3' end of the 2.85-kb mRNAprobes didnotdetect any antisense messages transcribed from this genomic map re-gion (data notshown). An antisenseoligonucleotide (primer 136) targeting the 3' end of the 2.3-kb cDNA could detect boththe 2.85-kb and 1.1-kbmessages. This result suggested that these two messages were transcribed in the same direction. Thetranscription pattern of this structurally tri-cistronic gene is very similar to that of HCMV tegument proteins pp65 and pp7l, in which two messages are tran-scribed from the corresponding genomic region. A large 4.0-kb mRNA and a small unspliced 1.9-kb mRNA are transcribed witha3'-coterminal end.The pp65 matrix tegu-ment protein is translated from a 4.0-kb mRNA, and the pp7l high-matrix tegument is made from a 1.9-kb mRNA (18).

Transcriptionkineticswerestudied with RNAs harvested at designated time points and after appropriate drug treat-ment. These studies revealed that the 2.85-kb and 1.1-kb mRNAs belong to two different classes of transcript. The 2.85-kb messageisanearly or ,-type mRNA which does not require virus DNA replication for maximal expression. The 1.1-kbmRNA,expressed late(48 h p.i.) after virus infection and requiring viral DNA replication for maximum expres-sion, isalateor y-type message.This differential expression occurs even thoughthese transcripts are derived from the samegenomic map region and overlap at the 3' end. This is common in the transcriptional regulation of early and late HCMVmRNAs,inwhichearly and late mRNAs are trans-lated afterimmediate-early mRNAs (2, 5, 22, 23).

VIROL.

on November 9, 2019 by guest

http://jvi.asm.org/

(11)

FIG. 8. Indirectimmunofluorescencestudy with antibodies againstHCMVfusion proteins. The rabbitantibodies used in this study were preabsorbedwith HEL cellular lysates. (Inset a)Mock-infectedHEL cells.(Ato D)HCMV-infected HEL cells at 12, 24, 48, and 72 hp.i., respectively,werereacted first with the antibody against the fusion protein and then withbiotin-conjugated goat anti-rabbit immunoglobulin Gantibody andstainedwithfluorescein-labeled avidin. Arrows indicate positive staining. Only results forUs18areshown. Panels a, A, and B weredeliberately overexposedtoshow cellmorphology.

The tricistronic 2.85-kb mRNA was transcribed after 12 h

p.i.andreachedthe maximumtranscriptionlevel at 24 h p.i. The first5'-end translationalproduct of the 2.85-kb mRNA, the

Us2O

polypeptide,wasdetectedat18 h p.i., and its level

increased slightly over the infection period. The

Us19

polypeptide couldbedetectedat averylowlevelat18 hp.i.,

buthigherlevels of

Us19

weredetectedbetween 24and 36 h

p.i. by immunofluorescence and Western blot analyses.

Transcription of the 2.85-kb mRNA decreased after it reached its maximum level. In contrast, translation of this message increased continuously over an extended period.

This result probably reflects the stability of the translation

products or an increase in the translation efficiency of the 2.85-kb mRNAatlate stagesof infection.

Aspointedoutabove,averyminor 1.8-kb mRNAspecies

wasinconsistently detectedwhenthe2.3-kb cDNAwasused as a probe in Northern hybridization. Several attempts to map transcription initiation by S1 mapping andprimer ex-tension failed. These results suggest that anextremelylow level oftranscription,if any, of the 1.8-kbmRNAmayoccur atthe later stages ofinfection. Thisspeculationis alsobased upon the isolation of a 1.8-kb cDNA, derived from this

region, from the cDNAlibrary ofHCMV-infected cells. At thistime,wedonothavedefinitive datatoconclude whether the 32-kDa

US19

peptideis derived fromthe 2.85-kborthe 1.8-kbtranscript. The32-kDa

Us19

peptidecouldbarelybe detected at18hp.i.butwasdetected atalevel

comparable

to that of the 43-kDa

Us20

peptide. If the stability and the

levelsoftranslation efficiencyarecomparablebetween these transcripts, then the level of mRNA for

Us19

should be

comparabletothatfor

Us20.

Immunofluorescentstaining of the

Us19

peptide indicated that it appeared as early as the

Us20

peptide,at24 hp.i.,butwaslessabundantthan

Us20.

From these data, we believe that the

Us19

ORF may be

translatedfromthe 2.85-kb mRNA.Expressionof this ORF mayrequire avirus-induced factor(s) for maximal activity.

The monocistronic 1.1-kb mRNAwas translated at a later timep.i.,andits translationsubsequentlyremainedconstant

throughout theinfection. This correlatedwellwith the late translation of the

US18

virus-specific polypeptide.

Antibodiesprepared againstthe threepeptides,expressed

as fusion proteins, detected virus-specific proteins in

HCMV-infected cell lysates. The differences between the sizes determined by Western blot analysis and the sizes

predictedfrom sequence analysis may be dueto posttrans-lational modification. These antibodies gavepositive

immu-nofluorescencewhen testedduringthe late stages of HCMV infection in HEL cells, but the virus neutralization assay resulted in lowneutralization activities. The low neutraliza-tionactivity of these antibodies could be dueto

inappropri-ate folding of the fusion proteins or the lack of posttrans-lational modification of these

virus-specific products

expressedin E. coli.It isalso

possible

thatthese

proteins

are

notessentialforinfectivity.

Expression

of these ORFswith VOL. 67, 1993

on November 9, 2019 by guest

http://jvi.asm.org/

[image:11.612.77.543.76.389.2]

(12)

HUANG

abaculovirusvectorin insect cells is underwayfor further functional and antigenic studies.

ACKNOWLEDGMENTS

WethankTimothyKowalik,BretWing,andDavidY.Huang for useful discussions andShu-mei Huongfortechnical assistance.

Thisinvestigationwassupported inpartbyPublicHealth Service grants AI-12717 and CA-21773 from the National Institutes of Health.

REFERENCES

1. Berk, A. J., and P. A. Sharp. 1977.Sizing and mapping ofearly

adenovirus mRNAsby gelelectrophoresis of S1 endonuclease-digested hybrids. Cell 12:721-732.

2. Chang,C. P., D. H. Vesole, J. Nelson, M. B.A. Oldstone, and M.F. Stinski. 1989. Identification and expression ofahuman cytomegalovirus earlyglycoprotein. J. Virol. 63:3330-3337. 3. Chee,M.S.,A. T.Bankier, S.Beck, R.Bohni, C.M.Brown,R.

Cerny,T.Horsnell, C.A. HutchisonIII, T. Kouzarides, J. A. Martignetti, E. Preddie, S. C. Satchwell, P.Tomlinson,K. M. Weston,and B.G.Barrell.1990.Analysisof theprotein-coding content of thesequence of HCMV strain AD169. Curr. Top. Microbiol. Immunol. 154:125-169.

4. Chou,P. Y., and G.D. Fasman. 1974. Prediction ofproteins conformation. Biochemistry 13:222-245.

5. Davis, M.G., andE.-S.Huang.1985.Nucleotidesequenceofa

humancytomegalovirus DNAfragment encoding a 67-kilodal-tonphosphorylated viralprotein. J. Virol. 56:7-11.

6. Gretch,D.R., B. Kari, R. Gehrz,and M. F. Stinski. 1988. A multigene family encodes the human cytomegalovirus glyco-protein complex gCII (gp47-52 complex). J. Virol. 62:1956-1962.

7. Harlow, E., and D. Lane.1988. Antibodies-alaboratory

man-ual, p. 90-91. Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

8. He, Y. S., L.Xu, and E.-S.Huang. 1992. Characterization of human cytomegalovirus UL84 earlygene andidentification of itsputative proteinproduct.J.Virol. 66:1098-1108.

9. Kilpatrick, B.A.,andE.-S.Huang. 1977. Human cytomegalo-virus genome: partial denaturation map and organization of genome sequences. J.Virol. 24:261-276.

10. Kingston, R. E. 1987. Guanidinium method for total RNA preparation,p.4.2.1.-4.5.3. InF. M.Ausubel,R. Brent,R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl(ed.),Currentprotocolsinmolecularbiology. John Wiley &Sons,New York.

11. Kingston, R. E. 1987.Primer extension,p.4.8.1.-4.8.3. In F. M. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J.A.Smith,and K.Struhl(ed.),Current protocols in molecularbiology.JohnWiley & Sons, New York.

12. Kozak, M. 1987. Ananalysis of 5'-noncoding sequences from 699 vertebratemessenger RNAs.Nucleic AcidsRes. 15:8125-8132.

13. LaFemina, R., and G. S.Hayward. 1980. Structuralorganization of theDNAmolecules from humancytomegalovirus,p.39-55. In R. Jaenisch, B. Fields, and C. F. Fox(ed.), Animal virus genetics. Academic Press,NewYork.

14. LaFemina,R.,andG. S.Hayward. 1983.Replicative forms of humancytomegalovirusDNAwithjoined terminiarefound in permissively infected human cells but not in non-permissive Balb/c-3T3mousecells. J. Gen. Virol. 64:373-389.

15. Mocarski, E. S.,L. Pereira,and L.McCormick.1988. Human cytomegalovirus ICP22, the product of the HWLF1 reading frame, isanearly nuclear protein that is released from cells.J. Gen. Virol. 69:2613-2621.

16. Nathans, J., and D. S. Hogness. 1983. Isolation, sequence analysis, and intron-exon arrangement of the gene encoding bovinerhodopsin. Cell 34:807-814.

17. Rowe, W. P., J. W. Hartley, S.W.Waterman, H.C.Turner, and R.J.Huebner. 1956. Cytopathogenicagentresembling human salivary gland virus recovered from tissue culture of human adenoids.Proc.Soc.Exp. Biol. Med. 92:418-424.

18. Ruger,B., S.Kiages,B.Walla, J.Albrecht,B.Fleckenstein,P. Tomlinson, and B. Barrell. 1987. Primarystructure and tran-scription of thegenescoding for thetwovirionphosphoproteins pp65andpp7l of human cytomegalovirus.J.Virol. 61:446-453. 19. Sanger,F., S.Nicklen,andA. R. Coulson. 1977. DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 75:5463-5467.

20. Seldon,R. F.1987.Analysisof DNA sequencesbyblottingand hybridization,p.2.9.1-2.9.9. In F. M.Ausubel,R.Brent,R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl(ed.),Currentprotocols in molecularbiology. John Wiley & Sons, New York.

21. Smith,D.B., and K. S. Johnson.1988.Single-steppurification of polypeptides expressedinEscherichia coliasfusionwith gluta-thioneS-transferase. Gene 67:31-40.

22. Stinski, M. F. 1978. Sequence of protein synthesis in cells infected by human cytomegalovirus: early and late virus-in-ducedpolypeptides. J.Virol. 26:686-701.

23. Wathen, M. W., and M. F. Stinski. 1982.Temporalpatternsof humancytomegalovirus transcription: mapping the viralRNAs synthesized at immediate-early, early, and late times after infection. J. Virol.41:462-477.

24. Weston, K. 1988. An enhancer element in the short unique region of human cytomegalovirus regulatesthe production of a group of abundant immediate early transcripts. Virology162: 406-416.

25. Weston, K., and B. G. Barrell. 1986. Sequence of the short unique region, short repeats and part of the long repeats of human cytomegalovirus. J. Mol. Biol. 192:177-208.